DOCSLIB.ORG

CART - Conversational Agent Research Toolkit Documentation Release 1.00

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
CART - Conversational Agent Research Toolkit Documentation Release 1.00 CART - Conversational Agent Research Toolkit Documentation Release 1.00 (Anonymized) May 06, 2021 Contents 1 Guide 1 1.1 About CART...............................................1 1.2 Installation................................................2 1.3 Using CART...............................................5 1.4 Modules................................................. 14 1.5 Tutorial.................................................. 14 1.6 License.................................................. 17 2 Indices and tables 19 i ii CHAPTER 1 Guide 1.1 About CART 1.1.1 Objectives The Conversational Agent Research Toolkit (CART) aims at enabling researchers to create conversational agents for experimental studies using computational methods. CART provides a unifying toolkit written in Python that integrates existing services and APIs for dialogue management, natural language understanding & generation, and frameworks that enable publishing the conversational agents as either a web interface or within messaging apps. Specifically developed for communication research, CART not only acts as an integration layer across these different services, but also aims to provides a configurable solution meeting the requirements of academic research. 1.1.2 Components CART acts as an integration layer between several services so it can generate a conversational agent that can interact with participants of an experiment, log the conversations, design experiments, and connect with questionnaires for self-reported measures. To do so, CART works with the following components: Dialogue Management CART currently uses DialogFlow as the primary tool to handle dialogue managemnt. This means that all the participant input, and the responses that the agent gives, are primarily setup in DialogFlow. Using specific tokens (see below), the researcher can customise the responses that the agent provides depending on the condition that the participant is in. Agent Publication CART currently uses the Microsoft Azure Bot Channel Registration to publish the conversational agent. Within CART, the Bot Framework is responsible for creating a webchat for users to chat with the agent. This webchat can be embedded in online surveys. The Bot Framework also allows the agent to be published in other channels (e.g., Skype, Telegram, Facebook Messenger) without needing to change the code within CART. 1 CART - Conversational Agent Research Toolkit Documentation, Release 1.00 Conversation Logging One of the key aspects of CART is the ability to log the conversations that participants in an experiment have with the agent. To do so, CART connects to a database under the control of the researcher. Experiment Design CART allows the researcher to create experiments in which the same agent acts in different ways depending on the condition that the participant is in. To do so, the researcher simply needs to add specific tokens in the Dialogue Management tool, and setup different conditions within CART itself. Online Questionnaire Integration After an agent is created and published, the researcher can integrate it to the flow of an online questionnaire (e.g., Qualtrics, SurveyMonkey). By turning on the questionnaire flow, the agent requests a unique ID from the participant, and returns a unique conversation code at the end of the conversation, so the researcher can link the conversation logs in the CART database with the responses in the online questionnaire. 1.2 Installation 1.2.1 Requirements Services and APIs CART acts as an integration layer between different services (and APIs), and specific configurations within CART itself. To use it, you will need: 1. Access to a SQL database 2. Access to a web service to publish CART (using Python 3.X) 3. An account with DialogFlow 4. An account with the Microsoft Bot Framework 5. A copy of the CART code Optionally, you also would need access to an online questionnaire tool (e.g., Qualtrics, SurveyMonkey) if you are integrating CART into a survey flow. Note: For #1 and #2, you can use your own server (if available), or AWS RDS (database) and Heroku (web service) as demonstrated in the Installation and Setup Guide. Python Packages CART uses a set of Python packages (e.g., PyMySQL, microsoftbotframework, google-cloud-dialogflow etc.). By downloading CART and running the step-by-step instructions in the installation guide (below), all the necessary pack- ages will be installed on the server. 2 Chapter 1. Guide CART - Conversational Agent Research Toolkit Documentation, Release 1.00 1.2.2 Installation and Setup Guide Important note: these steps show how to install and setup an agent using AWS RDS (as a database server) and Heroku (as a web service) along with DialogFlow and the MS Bot Framework. Advanced users can replace AWS RDS and Heroku by their own servers. Step 1. Download CART to your computer 1. Access [ANONYMIZED LINK OF THE GITHUB REPOSITORY] and select clone or download and download CART as a zip file. 2. Extract the zip file in a folder you want to store the agent Step 2. Rename the config.yaml file The config.yaml file, located at the folder cart, will contain all the basic configurations needed to connect to the services (MS Bot Framework and the SQL database), including usernames and passwords. Never make it publicly available. To create it, you need to: 1. Copy (or rename) the file called config_example.yaml to config.yaml 2. In the steps 3, 4 and 6 (below), you will need to update this file with information coming from each service. Step 3. Create an agent in DialogFlow 1. Log in to DialogFlow and select Create Agent. For CART, all tests have been done with the ES (simpler) version. 2. Follow the instructions from DialogFlow to enable the API, create a service account, and download the service key account file in JSON format. You will use it on step 5, as part of the web service configuration. Step 4. Create a database for logging CART requires the the URL (host) of a MySQL database, database name, and username and password to connect to and log the conversations between the agent and participants. The username provided needs to have all privileges (including CREATE) in the database. This information needs to be made included in the config.yaml file. If using AWS RDS, the following steps need to be followed: 1. Log into your AWS account and select RDS 2. Create a database using MySQL, and make sure to include the username and the password also in the config.yaml file 3. When asked for the database name, make sure to inform it, and also include it in the config.yaml file. 4. Select the endpoint of the database (check the dabatase details page), and paste it in the database_url field of config.yaml. 5. After the database is launched, make sure to check the security group (see database details), and open the inbound port to the server that you will be using. Notes: 1.2. Installation 3 CART - Conversational Agent Research Toolkit Documentation, Release 1.00 • Usually a micro or small instance in AWS is sufficient for testing purposes, and can later be upgraded to medium/large when the agent is live and handling several conversations at the same time (e.g., during an experiment) • For security reasons, it is recommended not to allow any IP to connect to your database. When using Heroku, see this list of plugins that can create a static IP for your app. • When running the agent in the server for the first time, it will automatically create two tables in the database: – logs, which records every interaction between the agent and the participant – conversations, which records each individual conversation (i.e., in general, each participant = one conversation) that is started with the agent. Step 5. Publish the agent as a web service After completing the steps above, it is time to publish the agent as a web service. This can be done using Heroku (as demonstrated below) or in any other server that supports Python and Flask applications. It is important to note that the server should also be able to serve pages in https. If using Heroku: 1. Log into your Heroku account and create a new app 2. Set the environment variables (called in Heroku “config vars”) for DialogFlow. See note 2, below, for details. 3. Select the deployment method 4. Deploy the app 5. After the build has been completed, select open app 6. Copy the URL of the app (it should start with the app-name, and end with herokuapp.com) to use in the next step Note 1: the URL of the app is needed so it can be registered in the MS Bot Framework (next step). The registration in the MS Bot Framework will also provide authentication credentials. These credentials will need to be added to the config.yaml file, and the agent will need to be published again in Heroku (as outlined in the next step). Note 2: Open the service account file downloaded on step 3(above) locally in a text editor, remove all line breaks, and substitute the double quotes (”) by single quotes (‘). In the web service, add this as an environment variable called DF_CREDENTIALS. Create another environment variable called DF_LANGUAGE_CODE and set its value to the appropriate language (e.g., en). Step 6. Connect the agent to the MS Bot Framework After a URL for the agent as a web service is available (e.g., for Heroku: https://NAMEOFTHEAPP. herokuapp.com/), the agent can be registered in the MS Bot Framework. To do so: 1. Log in your Microsoft Bot Framework account, selecting My Bots 2. Select create a bot. You will be redirected to Azure Bot Service 3. Select Bot Channels Registration 4. Provide the information required 5. The messaging endpoint will be the URL of the Heroku app + api/messages - example: https:// NAMEOFTHEAPP.herokuapp.com//api/messages 4 Chapter 1. Guide CART - Conversational Agent Research Toolkit Documentation, Release 1.00 6.
Load more

Recommended publications
  • Google Docs Accessibility (Pdf)
    Google Docs Accessibility (A11y) Building Accessible Google Docs • Heading Styles • Images • Table of Contents • Captioning • Columns and Lists • Tables A11y • Tab Stops • Color Contrast • Paragraph Spacing • Headers and Footers • Meaningful Link Text • Accessibility Checker What is Assistive Technology? Assistive Technology (AT) are “products, equipment, and systems that enhance learning, working, and daily living for persons with disabilities.” Magnification Speech Screen Readers Software Recognition Trackball Mouse Keyboard Zoom Text Braille Computer Keyboard Captions/Subtitles Captioned Telephone Video Relay Services Captioning Videos Per federal and state law, and CSU policy, instructional media (e.g., videos, captured lectures, recorded presentations) must have captions. This includes instructional media used in classrooms, posted on websites or shared in Canvas. • All students who are enrolled in a course must be able to access the content in the course. • Faculty: Funding is available to help faculty generate captions and transcripts for instructional media. Materials should be submitted at least six weeks in advance of their use in instruction. • Staff: For CSUN staff who do not provide classroom material, there is a cost through chargeback. For information on the chargeback, email [email protected]. csun.edu/captioning What are Screen Readers Screen readers are a form of assistive technology (AT) software that enables access to a computer, and all the things a computer does, by attempting to identify and interpret what is being displayed on the computer screen using text-to-speech. Screen readers can only access and process live text (fully editable or selectable text). • Provides access to someone who is visually impaired, mobility or has a learning disability to access text on the screen.
    [Show full text]
  • Profiles Research Networking Software Installation Guide
    Profiles Research Networking Software Installation Guide Documentation Version : July 25, 2014 Software Version : ProfilesRNS_2.1.0 Table of Contents Introduction ..................................................................................................................... 2 Hardware and Operating System Requirements ............................................................. 3 Download Options ........................................................................................................... 4 Installing the Database .................................................................................................... 5 Loading Person Data....................................................................................................... 8 Loading Person Data: Part 1 – Importing SSIS Packages into SQL Server msdb Database ..................................................................................................................... 8 Loading Person Data: Part 2 – Importing Demographic Data .................................... 10 Loading Person Data: Part 3 – Geocoding ................................................................ 15 Loading Person Data: Part 4 – Obtaining Publications .............................................. 16 Loading Person Data: Part 5 – Convert data to RDF ................................................. 19 Scheduling Database Jobs ............................................................................................ 21 Installing the Code........................................................................................................
    [Show full text]
  • A Comparison of Natural Language Understanding Platforms for Chatbots in Software Engineering
    1 A Comparison of Natural Language Understanding Platforms for Chatbots in Software Engineering Ahmad Abdellatif, Khaled Badran, Diego Elias Costa, and Emad Shihab, Senior Member, IEEE Abstract—Chatbots are envisioned to dramatically change the future of Software Engineering, allowing practitioners to chat and inquire about their software projects and interact with different services using natural language. At the heart of every chatbot is a Natural Language Understanding (NLU) component that enables the chatbot to understand natural language input. Recently, many NLU platforms were provided to serve as an off-the-shelf NLU component for chatbots, however, selecting the best NLU for Software Engineering chatbots remains an open challenge. Therefore, in this paper, we evaluate four of the most commonly used NLUs, namely IBM Watson, Google Dialogflow, Rasa, and Microsoft LUIS to shed light on which NLU should be used in Software Engineering based chatbots. Specifically, we examine the NLUs’ performance in classifying intents, confidence scores stability, and extracting entities. To evaluate the NLUs, we use two datasets that reflect two common tasks performed by Software Engineering practitioners, 1) the task of chatting with the chatbot to ask questions about software repositories 2) the task of asking development questions on Q&A forums (e.g., Stack Overflow). According to our findings, IBM Watson is the best performing NLU when considering the three aspects (intents classification, confidence scores, and entity extraction). However, the results from each individual aspect show that, in intents classification, IBM Watson performs the best with an F1-measure>84%, but in confidence scores, Rasa comes on top with a median confidence score higher than 0.91.
    [Show full text]
  • Connected Cl As Sroom
    Data in the Cloud he ability to move from one representation monthly temperatures for cities in the United of data to another is one of the key char- States, South America, and Russia. Tacteristics of expert mathematicians and A quick inspection of the graph makes it evi- scientists. Cloud computing will offer more dent that on average it is always colder in Verk- opportunities to create and display multiple hoyansk than in Washington (and much colder representations of data, making this skill in the winter) and that the seasons in the south- even more important in the future. ern hemisphere are reversed. These patterns would have been much more difficult to discern Multiple Representations in a table of numbers alone. We can represent data in a variety of forms— Even young students can explore multiple rep- graphs, charts, tables of numbers, equations. resentations of the same data with software such Mathematicians, scientists, and engineers often as The Graph Club (Tom Snyder Productions). look for patterns in data. Different representations This application allows students to view two of the same data sometimes make it easier to see linked representations of data simultaneously. As a pattern. For example, the pattern in the table children drag icons to form a picture graph or of numbers below is not immediately evident. enter numbers in a simple table, a corresponding bar graph or pie chart takes shape (see Figure 2). Changing the size of a sector in the circle graph or the height of a bar in a bar chart Table 1. changes the pictogram or vice versa.
    [Show full text]
  • Voice.AI Gateway One-Click Dialogflow Integration Guide
    Integration Guide AudioCodes Intuitive Human Communications for Chatbot Services Voice.AI Gateway Google One-Click DialogFlow Integration Integration Guide Contents Table of Contents 1 Introduction ......................................................................................................... 7 1.1 Purpose .................................................................................................................. 7 2 Messages Sent by Voice.AI Gateway ................................................................ 9 2.1 Welcome Message ................................................................................................. 9 3 Messages Sent by Agent .................................................................................. 11 3.1 Basic Activity Syntax ............................................................................................ 11 3.2 hangup activity ..................................................................................................... 11 3.3 Bot Framework Specific Details ............................................................................ 12 3.3.1 Google Dialogflow ....................................................................................................12 Voice Bot Solutions 3 Voice.AI Gateway One-Click Dialogflow List of Tables Table 2-1: Description of Initial Message Sent by Voice.AI Gateway ...................................................... 9 Table 3-1: Properties of JSON Object Activities.....................................................................................11
    [Show full text]
  • The Ultimate Guide to Google Sheets Everything You Need to Build Powerful Spreadsheet Workflows in Google Sheets
    The Ultimate Guide to Google Sheets Everything you need to build powerful spreadsheet workflows in Google Sheets. Zapier © 2016 Zapier Inc. Tweet This Book! Please help Zapier by spreading the word about this book on Twitter! The suggested tweet for this book is: Learn everything you need to become a spreadsheet expert with @zapier’s Ultimate Guide to Google Sheets: http://zpr.io/uBw4 It’s easy enough to list your expenses in a spreadsheet, use =sum(A1:A20) to see how much you spent, and add a graph to compare your expenses. It’s also easy to use a spreadsheet to deeply analyze your numbers, assist in research, and automate your work—but it seems a lot more tricky. Google Sheets, the free spreadsheet companion app to Google Docs, is a great tool to start out with spreadsheets. It’s free, easy to use, comes packed with hundreds of functions and the core tools you need, and lets you share spreadsheets and collaborate on them with others. But where do you start if you’ve never used a spreadsheet—or if you’re a spreadsheet professional, where do you dig in to create advanced workflows and build macros to automate your work? Here’s the guide for you. We’ll take you from beginner to expert, show you how to get started with spreadsheets, create advanced spreadsheet-powered dashboard, use spreadsheets for more than numbers, build powerful macros to automate your work, and more. You’ll also find tutorials on Google Sheets’ unique features that are only possible in an online spreadsheet, like built-in forms and survey tools and add-ons that can pull in research from the web or send emails right from your spreadsheet.
    [Show full text]
  • Recovering Semantics of Tables on the Web
    Recovering Semantics of Tables on the Web Petros Venetis Alon Halevy Jayant Madhavan Marius Pas¸ca Stanford University Google Inc. Google Inc. Google Inc. [email protected] [email protected] [email protected] [email protected] Warren Shen Fei Wu Gengxin Miao Chung Wu Google Inc. Google Inc. UC Santa Barbara Google Inc. [email protected] [email protected] [email protected] [email protected] ABSTRACT away. Furthermore, unlike text documents, where small changes The Web offers a corpus of over 100 million tables [6], but the in the document structure or wording do not correspond to vastly meaning of each table is rarely explicit from the table itself. Header different content, variations in table layout or terminology change rows exist in few cases and even when they do, the attribute names the semantics significantly. In addition to table search, knowing are typically useless. We describe a system that attempts to recover the semantics of tables is also necessary for higher-level operations the semantics of tables by enriching the table with additional anno- such as combining tables via join or union. tations. Our annotations facilitate operations such as searching for In principle, we would like to associate semantics with each table tables and finding related tables. in the corpus, and use the semantics to guide retrieval, ranking and To recover semantics of tables, we leverage a database of class table combination. However, given the scale, breadth and hetero- labels and relationships automatically extracted from the Web. The geneity of the tables on the Web, we cannot rely on hand-coded database of classes and relationships has very wide coverage, but domain knowledge.
    [Show full text]
  • A Simple Generic Attack on Text Captchas
    A Simple Generic Attack on Text Captchas Haichang Gao1*, Jeff Yan2*, Fang Cao1, Zhengya Zhang1, Lei Lei1, Mengyun Tang1, Ping Zhang1, Xin Zhou1, Xuqin Wang1 and Jiawei Li1 1. Institute of Software Engineering, Xidian University, Xi’an, Shaanxi, 710071, P.R. China 2. Security Lancaster & School of Computing and Communications, Lancaster University, UK ∗Corresponding authors: [email protected], [email protected] Abstract—Text-based Captchas have been widely deployed attacked many early Captchas deployed on the Internet [19]. across the Internet to defend against undesirable or malicious Yan and El Ahmad broke most visual schemes provided at bot programs. Many attacks have been proposed; these fine prior Captchaservice.org in 2006 [24], published a segmentation art advanced the scientific understanding of Captcha robustness, attack on Captchas deployed by Microsoft and Yahoo! [25] but most of them have a limited applicability. In this paper, in 2008, and broke the Megaupload scheme with a method we report a simple, low-cost but powerful attack that effectively of identifying and merging character components in 2010 [1]. breaks a wide range of text Captchas with distinct design features, including those deployed by Google, Microsoft, Yahoo!, Amazon In 2011, Bursztein et al. showed that 13 Captchas on pop- and other Internet giants. For all the schemes, our attack achieved ular websites were vulnerable to automated attacks, but they a success rate ranging from 5% to 77%, and achieved an achieved zero success on harder schemes such as reCAPTCHA average speed of solving a puzzle in less than 15 seconds on and Google’s own scheme [5].
    [Show full text]
  • Webkit and Blink: Open Development Powering the HTML5 Revolution
    WebKit and Blink: Open Development Powering the HTML5 Revolution Juan J. Sánchez LinuxCon 2013, New Orleans Myself, Igalia and WebKit Co-founder, member of the WebKit/Blink/Browsers team Igalia is an open source consultancy founded in 2001 Igalia is Top 5 contributor to upstream WebKit/Blink Working with many industry actors: tablets, phones, smart tv, set-top boxes, IVI and home automation. WebKit and Blink Juan J. Sánchez Outline The WebKit technology: goals, features, architecture, code structure, ports, webkit2, ongoing work The WebKit community: contributors, committers, reviewers, tools, events How to contribute to WebKit: bugfixing, features, new ports Blink: history, motivations for the fork, differences, status and impact in the WebKit community WebKit and Blink Juan J. Sánchez WebKit: The technology WebKit and Blink Juan J. Sánchez The WebKit project Web rendering engine (HTML, JavaScript, CSS...) The engine is the product Started as a fork of KHTML and KJS in 2001 Open Source since 2005 Among other things, it’s useful for: Web browsers Using web technologies for UI development WebKit and Blink Juan J. Sánchez Goals of the project Web Content Engine: HTML, CSS, JavaScript, DOM Open Source: BSD-style and LGPL licenses Compatibility: regression testing Standards Compliance Stability Performance Security Portability: desktop, mobile, embedded... Usability Hackability WebKit and Blink Juan J. Sánchez Goals of the project NON-goals: “It’s an engine, not a browser” “It’s an engineering project not a science project” “It’s not a bundle of maximally general and reusable code” “It’s not the solution to every problem” http://www.webkit.org/projects/goals.html WebKit and Blink Juan J.
    [Show full text]
  • Webkit and Blink: Bridging the Gap Between the Kernel and the HTML5 Revolution
    WebKit and Blink: Bridging the Gap Between the Kernel and the HTML5 Revolution Juan J. Sánchez LinuxCon Japan 2014, Tokyo Myself, Igalia and WebKit Co-founder, member of the WebKit/Blink/Browsers team Igalia is an open source consultancy founded in 2001 Igalia is Top 5 contributor to upstream WebKit/Blink Working with many industry actors: tablets, phones, smart tv, set-top boxes, IVI and home automation. WebKit and Blink Juan J. Sánchez Outline 1 Why this all matters 2 2004-2013: WebKit, a historical perspective 2.1. The technology: goals, features, architecture, ports, webkit2, code, licenses 2.2. The community: kinds of contributors and contributions, tools, events 3 April 2013. The creation of Blink: history, motivations for the fork, differences and impact in the WebKit community 4 2013-2014: Current status of both projects, future perspectives and conclusions WebKit and Blink Juan J. Sánchez PART 1: Why this all matters WebKit and Blink Juan J. Sánchez Why this all matters Long time trying to use Web technologies to replace native totally or partially Challenge enabled by new HTML5 features and improved performance Open Source is key for innovation in the field Mozilla focusing on the browser WebKit and now Blink are key projects for those building platforms and/or browsers WebKit and Blink Juan J. Sánchez PART 2: 2004-2013 WebKit, a historical perspective WebKit and Blink Juan J. Sánchez PART 2.1 WebKit: the technology WebKit and Blink Juan J. Sánchez The WebKit project Web rendering engine (HTML, JavaScript, CSS...) The engine is the product Started as a fork of KHTML and KJS in 2001 Open Source since 2005 Among other things, it’s useful for: Web browsers Using web technologies for UI development WebKit and Blink Juan J.
    [Show full text]
  • Ate: a Map - Fusion Tables Help
    2/19/2014 Create: a map - Fusion Tables Help Help Community Fusion Tables is an experimental app. Create: a map Use Fusion Tables' new look with this tutorial This 2-3 minute tutorial steps you through uploading data to generate a map with placemarks in the new version of Fusion Tables. Tutorial Contents: Import the data Map the data Customize the info window template [optional] For this tutorial, a sample dataset is provided containing records of insects collected in Costa Rica by ecologists. Each location is recorded with a latitude and longitude point. https://support.google.com/fusiontables/answer/2527132?hl=en&topic=2573107&ctx=topic#importsample 1/5 2/19/2014 Create: a map - Fusion Tables Help Import the sample dataset 1. Download this sample insect dataset. 2. Go to Google Docs. Sign in to your Google Account or create a Google Account if you don't already have one. (Note that you can't use a Google Apps for your Domain account for Fusion Tables.) 3. Click the Create button. 4. Choose Create > More > Fusion Table from the dropdown menu. 5. In the Import new table dialog box, click Choose File. 6. Select the "Astraptes fulgerator complex sample data.csv" file you downloaded, and press Next. 7. Check that the data is formatted correctly and click Next. 8. Give your table a name and click Finish. Your uploaded data now appears in a new Fusion Table with thumbnail images in the two columns of URL links: Map the data Fusion Tables auto-detects location data in a table and displays a tab called "Map of <location column name>." In this case, the Map tab is titled "Map of latitude." 1.
    [Show full text]
  • Large Scale Captcha Survey
    LARGE SCALE CAPTCHA SURVEY by Mecheal Greene A thesis submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Master of Science in Electrical and Computer Engineering Summer 2018 c 2018 Mecheal Greene All Rights Reserved LARGE SCALE CAPTCHA SURVEY by Mecheal Greene Approved: Haining Wang, Ph.D. Professor in charge of thesis on behalf of the Advisory Committee Approved: Kenneth E. Barner, Ph.D. Chair of the Department of Electrical and Computer Engineering Approved: Babatunde A. Ogunnaike, Ph.D. Dean of the College of Engineering Approved: Douglas J. Doren, Ph.D. Interim Vice Provost for Graduate and Professional Education ACKNOWLEDGMENTS First, I would like to thank the Bridge to Doctorate program for funding me and allowing me the opportunity to pursue my Masters degree at the University of Delaware, immediately following my graduation from Cheyney University. I especially want to thank Dean Doctor Michael Vaughan for being the on campus advisor of the program and assisting me whenever a problem arose. My advisor, Doctor Haining Wang, for guiding me on the completion of my Masters thesis and degree. You allowed me to grow and learn throughout my graduate career. Most, importantly you took a chance on a young black man that you didnt even know but still you embraced me with open arms and worked with me towards accomplishing this goal. My partner on this project Alparslan Sari, you helped me and guided me every step of the way whether it was through meetings, conversations or emails. You knew this was my first time doing anything thesis related but you made it all easier on me and didnt mind helping me above and beyond what I even imagined.
    [Show full text]